1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
const axios = require('axios');
const api_key = "YOUR API-KEY";
const url = "https://api.segmind.com/v1/veed-video-background-removal";
const data = {
"video_url": "https://segmind-resources.s3.amazonaws.com/input/veed-video-background-removal-input.mp4",
"mode": "standard"
};
(async function() {
try {
const response = await axios.post(url, data, { headers: { 'x-api-key': api_key } });
console.log(response.data);
} catch (error) {
console.error('Error:', error.response.data);
}
})();Attributes
The source video to remove the background from. Provide a clear, well-lit HD clip (mp4, mov, webm, m4v, gif) for the cleanest matte.
Chooses the matting engine. Use standard for best quality, fast for cheap high-volume jobs, green-screen for chromakey footage.
Allowed values:
Sharpens hair and fine edges at a higher per-frame rate. Keep on for hero content, off in fast mode to save cost. Standard/fast only.
Optimizes matting for the subject type. Leave true for people; set false for products and other objects. Standard/fast only.
Sets the output container. vp9 gives one WebM with alpha; h264 returns separate RGB + alpha URLs for higher RGB quality.
Allowed values:
Controls green-spill cleanup in green-screen mode only. Raise toward 1 if green remains; lower if subject colors shift.
min : 0,
max : 1
To keep track of your credit usage, you can inspect the response headers of each API call. The x-remaining-credits property will indicate the number of remaining credits in your account. Ensure you monitor this value to avoid any disruptions in your API usage.
VEED Video Background Removal — AI Video Matting API
What is VEED Video Background Removal?
VEED Video Background Removal is an AI video matting API that separates subjects from their backgrounds in any video — no green screen required. Instead of frame-by-frame rotoscoping, the model analyzes pixel data, edges, and motion directly to produce a clean per-frame alpha matte for people and objects. Delivered through Segmind on fal.ai infrastructure, it turns ordinary talking-head clips, product shots, and webcam recordings into transparent, composite-ready footage in a single API call. You send a video URL, choose a mode, and receive an alpha-enabled result you can drop onto brand colors, custom images, or AI-generated scenes.
Key Features
- No green screen needed — automatic subject extraction from natural, cluttered, or moving backgrounds.
- Three modes —
standardfor best quality,fastfor cheaper high-volume jobs, andgreen-screenfor chromakey footage with green-spill suppression. - Edge refinement — optional
refine_foreground_edgessharpens hair and fine details at a higher per-frame rate. - Subject optimization —
subject_is_persontunes matting for human subjects versus objects. - Flexible output — VP9 gives a single WebM with an embedded alpha channel; H264 returns separate RGB + alpha videos for higher RGB quality in post.
- Broad input support — MP4, MOV, WebM, M4V, and GIF via URL.
Best Use Cases
Use it to produce polished talking-head videos for social, YouTube, and course content without a studio. It shines for product demos and e-commerce clips where subjects need to sit cleanly on brand backgrounds, for webcam and screen-recording overlays, and for compositing pipelines that would otherwise require manual masking. Green-screen mode is ideal when you already shot on a chroma backdrop and want automatic keying with spill cleanup. Fast mode is built for high-volume, cost-sensitive batch workflows where slight edge artifacts are acceptable.
Prompt Tips and Output Quality
This is a parameter-driven model, not a prompt-driven one — quality comes from picking the right mode and toggles. Feed HD video where the subject is well-lit and clearly separated from the background for the cleanest mattes. Keep refine_foreground_edges on for hero content with visible hair or fine edges; turn it off in fast mode to cut cost on bulk jobs. Set subject_is_person true for people and false for objects. Choose VP9 for a simple single-file alpha workflow, or H264 when you need maximum RGB fidelity. In green-screen mode, raise spill_suppression_strength toward 1 if green fringing remains, or lower it if the subject's colors start to shift.
FAQs
Do I need a green screen? No. Standard and fast modes remove backgrounds from any footage. Green-screen mode is a separate option for people who already shot on chroma.
What output formats does it return? VP9 (default) is a single WebM with an alpha channel; H264 returns JSON with separate RGB and alpha video URLs.
What video formats can I upload? MP4, MOV, WebM, M4V, and GIF, provided as a URL.
What's the difference between standard and fast? Standard prioritizes matte quality; fast trades a little edge precision for lower cost and quicker turnaround on large batches.
When should I turn on edge refinement? For close-ups and subjects with hair or intricate outlines where clean edges matter most — it improves quality at a higher per-frame rate.
Does it work for objects, not just people? Yes. Set subject_is_person to false to optimize matting for non-human subjects.
Popular Models
